- Email: [email protected]
Indocyanine green angiography in birdshot chorioretinopathy
Indocyanine Green Angiography in Birdshot Chorioretinopathy Christine Fardeau, MD,1 Carl P. Herbort, MD, PD,2,3 Nathalie Kullmann, MD,1 Gabriel Quentel, MD,4 Phuc LeHoang, MD, PhD1 Objective: Birdshot chorioretinopathy (BC) is an ocular inflammatory disease involving both the retina and the choroid. The study goal was to evaluate indocyanine green angiographic features in BC to assess choroidal involvement. Design: Retrospective, observational case series. Participants: Fifty-two patients with BC documented with at least 1 concomitant fluorescein and indocyanine green angiogram. Intervention: Indocyanine green angiography (ICGA) was performed according to a standard protocol used for inflammatory disorders. Main Outcome Measure: Indocyanine green angiographic signs were correlated with fundus photographs, fluorescein angiography, degree of inflammatory activity, and stage of disease. Results: In active disease, three main features were observed. The principal finding, found in 100% of patients, was the presence of hypofluorescent dark dots during the intermediate phase of angiography; their evolutionary pattern was twofold, becoming either isofluorescent or remaining hypofluorescent at the late phase of angiography. The other two signs were fuzzy, indistinct choroidal vessels and late-diffuse choroidal hyperfluorescence. In chronic longlasting disease, the characteristic finding was the presence of hypofluorescent dark dots that persisted in the late phase of disease and is theorized to correspond either to chorioretinal atrophy (irregular geographic pattern) or to persistent choroidal granulomas (round oval form). Conclusions: Consistent ICGA findings in 52 patients allowed the authors to establish a fairly precise ICGA semiology for BC. This procedure enabled the authors to assess choroidal involvement, and, in selected cases, it also was found to be of diagnostic help and useful to monitor therapeutic intervention. Ophthalmology 1999;106:1928 –1934 Birdshot chorioretinopathy (BC) is characterized by deep cream-colored dots scattered diffusely throughout both fundi. Other signs are retinal vasculitis, papillitis, cystoid macular edema, electroretinographic changes, and potentially severe visual field defects.1– 4 The course of the disease is characterized by recrudescence of uveitis, which may progressively result in reduced visual function due to cystoid macular edema, macular atrophy, or, less frequently, due to subretinal neovessels. Fluorescein angiography (FA) is useful to determine the extent of retinal involvement, which includes retinal vasculitis causing occasional patchy hyperfluorescence, disc hyperfluorescence, cystoid or difOriginally received: November 16, 1998. Revision accepted: June 1, 1999. Manuscript no. 98744. 1 Department of Ophthalmology, Centre Hopitalo-Universitaire Pitie´-Salpe´trie`re, Paris, France. 2 Department of Ophthalmology, Hoˆpital Jules Gonin, Lausanne, Switzerland. 3 La Source Eye Center, Lausanne, Switzerland. 4 Centre d’Imagerie et de Laser, Paris, France. Address correspondence to Carl P. Herbort, MD, PD, La Source Eye Center, 2 Avenue des Bergie`res, CH-1004 Lausanne, Switzerland. Reprint requests to Phuc LeHoang, MD, PhD, Department of Ophthalmology, Centre Hopitalo-Universitaire Pitie´-Salpe´trie`re, 83 Boulevard de l’Hoˆpital, F-75013 Paris, France.
1928
fuse macular edema, and pseudodelay in arteriovenous fluorescein circulation time.5,6 Inflammatory activity usually is prominent equally in the retina and the choroid.1,2 To assess involvement of the latter, indocyanine green angiography (ICGA) may prove to be very useful as it has in other posterior uveitic conditions involving the choroid.7–9 The absorption peak of indocyanine green (ICG) is situated at ⫾795 nm, and emission of fluorescent light occurs at approximately 830 nm (nearinfrared wavelengths), thus allowing visualization of choroidal fluorescence through the pigmentary epithelium.10 In contrast to fluorescein, ICG is not only highly proteinbound (98%) but binds up to 80% to larger proteins (globulins and alpha-1-lipoproteins). Accordingly, ICG does not leak from normal or only slightly inflamed retinal vessels but leaks unimpaired, albeit slowly, from the fenestrated choriocapillaris, progressively impregnating the choroidal space.11 Because of its particular fluorescence parameters, ICG provides visual access to the choroid and appears to be especially useful to evaluate choroidal involvement of inflammatory disorders.12 The aim of this study was to determine ICGA signs in BC in correlation with clinical signs, disease stage, fundus photographs, and FA findings. We also investigated the possible use of ICGA to monitor both disease progression and effect of immunosuppressive therapy.
Fardeau et al 䡠 ICG Angiography in Birdshot Chorioretinopathy
Patients and Methods Charts of patients with the diagnosis of BC seen in the Department of Ophthalmology of Hoˆpital de la Pitie´-Salpe´trie`re in Paris, the Department of Ophthalmology at Hoˆpital Jules Gonin in Lausanne, and at La Source Eye Center in Lausanne were analyzed.
Inclusion Criteria Patients for whom the diagnosis of BC was verified and who had at least one ICG and fluorescein angiogram performed were included in the study. The diagnosis of BC was based on the clinical and angiographic criteria described in the introduction and by the presence of the HLA-A29 antigen. When necessary, investigations were performed to exclude other possible conditions that could be mistaken for BC, such as sarcoidosis, syphilis, and tuberculosis. Typically, this comprised all or part of the following: complete blood count, erythrocyte sedimentation rate, serum angiotensinconverting enzyme and lysozyme, herpes virus and syphilis serologies, search for autoantibodies and anticardiolipin antibodies, chest x-ray, purified protein derivative skin test, and, occasionally, gallium scan and biopsy of salivary glands. Multifocal choroiditis and multiple evanescent white dot syndrome were excluded on the bases of clinical grounds and angiographic findings.
Treatment The mainstay of therapy consisted of systemic corticosteroid therapy alone in 8 patients, of intravenous immunoglobulins alone in 5 patients, of corticosteroid therapy ⫹ cyclosporin A in 17 patients, and of corticosteroid and/or cyclosporin A and intravenous immunoglobulins in 11 patients; no treatment was given in 11 patients.
Angiographic Procedure A Topcon 1A conventional fundus camera coupled either to an OIS (Opthalmic Imaging System, Sacramento, CA) or an ImageNet (Topcon, Tokyo, Japan) image digitalizing system was used. Conventional fundus camera systems rather than scanning laser ophthalmoscopic systems should be used for uveitis workup because they allow easy panoramic imaging of the whole fundus as well as good reproducibility of angiograms for follow-up examinations and provide good quality late-phase ICG pictures, three very important elements in the angiographic investigation of uveitis. In contrast, the disadvantages of scanning laser ophthalmoscopic videoangiography are partial nonsystematic examination of the fundus, poor visualization of the periphery, and low-quality late ICG pictures. This confocal system allows precise measurement of fluorescence in a given plane of focalization and is therefore very useful when the plane of analysis is well-determined such as in subretinal neovascular membranes. In uveitis, however, lesions may be located from the retinal plane to the scleral plane. Therefore, the global superimposed fluorescence obtained with conventional fundus camera digital angiography is more useful than the focalized images provided by scanning laser ophthalmoscopic systems, which are more difficult to standardize and interpret in routine practice. To exclude autofluorescence, preinjection fluorescence was looked for with the highest flash intensity used for ICGA. At the same time, red-free posterior pole frames were taken. Fluorescein angiography and ICGA usually were performed during the same
session. The procedure typically was initiated with a bolus injection of 50 mg of indocyanine green (Cardiogreen; Becton Dickinson, Cockeysville, MD) diluted in 7.5 ml of a 0.9% salt solution. Early ICGA frames of the posterior pole were taken (to approximately 2–3 minutes; early phase of angiography). Indocyanine green background fluorescence was analyzed in detail at 10 ⫾ 3 minutes after ICG injection (intermediate phase of angiography) by taking frames of the posterior pole and the periphery (a minimum of eight frames of the entire periphery for 360°). At the end of the intermediate phase of ICGA, FA was performed. Early FA frames of the posterior pole were taken up to 2 minutes, followed by 360° panorama imaging of the periphery and late fluorescein posterior pole frames at 10 minutes. Late ICG background fluorescence was analyzed at 35 ⫾ 10 minutes after ICG injection (late phase of angiography) in the same fashion as during the intermediate phase.13
Image Analysis The surface area of hypofluorescent dots in the intermediate phase of ICG angiograms was compared to that of red-free photographs by mapping after computerized overlay. Indocyanine green angiographic findings were related to clinical data, fundus color, green and blue light photographs, and FA findings. The patient history, disease stage, and treatment regimen were other factors considered. In cases with more than one ICG angiogram, evolution of angiographic signs were correlated with clinical evolution and extent of treatment.
Results A total of 52 patients (27 women and 25 men; mean age, 49 ⫾ 7 years at presentation) were included in the study. Twenty-two patients were seen in an active phase of disease, which included both recurrences in insufficiently treated patients and in newly diagnosed patients before treatment. Within this group, both pretreatment and post-treatment angiograms were available on 16 of 22 patients. Thirty patients had their ICG angiograms taken “offtreatment” in a burn-out phase of disease or while they were receiving therapy that was considered to control disease activity. Most of the 52 patients had roughly symmetric clinical involvement, the number of cream-colored areas varying from a few lesions to widespread involvement from the posterior pole to the equator.
Fluorescein Angiography In the active phase of disease, FA hyperfluorescence of the optic disc was present bilaterally in all patients (n ⫽ 22). Diffuse bilateral edema of the posterior pole caused by leakage, particularly along the temporal vascular arcades, was also present in all patients and had a cystoid component in 16 patients (Fig 1). Vasculitis of small retinal vessels was documented in 20 patients, appearing as small FA hyperfluorescent foci scattered over the entire fundus giving a mottled pattern (Figs 1 and 2). These FA hyperfluorescent foci, becoming progressively more fluorescent on late frames of the FA angiogram, indicated retinal inflammatory involvement (Fig 2) because they did not correspond to the creamcolored lesions seen clinically (not shown). The cream-colored fundus lesions were silent (or relatively hypofluorescent) on FA. In seven patients, early FA frames were sufficient to document
1929
Ophthalmology Volume 106, Number 10, October 1999
Figure 2. Dual choroidal and retinal involvement. Newly diagnosed birdshot chorioretinopathy. On fluorescein angiography (FA), the characteristic patchy, mottled hyperfluorescence secondary to vasculitis of small vessels is scattered throughout the fundus. Comparative analysis of indocyanine green angiography (ICGA) and FA shows that the patchy fluorescein hyperfluorescence does not correspond to hypofluorescent areas on ICGA. The latter areas are either silent or hypofluorescent on FA.
Figure 1. Hypofluorescent dots. Birdshot chorioretinopathy with a subacute course of 5 years duration. Fluorescein angiography (top pictures) shows disc and diffuse posterior pole hyperfluorescence (edema). In the periphery (top right), there is typical mottled, patchy hyperfluorescence secondary to vasculitis of small vessels. Indocyanine green angiography (ICGA) shows typical hypofluorescent areas well-apparent in the intermediate phase (middle pictures). Most of the hypofluorescent areas still are present in the late phase (bottom pictures), although some have disappeared in the late phase. Note also diffuse late-phase choroidal hyperfluorescence and fuzziness of choroidal vessels in the intermediate phase (middle pictures). This patient had been followed for 5 years for a uveitis of undetermined origin. Typical pale areas were not present on fundus examination because of an albinotic fundus. Indocyanine green angiography was helpful in making the correct diagnosis.
pseudodelay in retinal arteriovenous fluorescein circulation time, as described by Gass,5 which was not found by ICGA as described earlier6 (Fig 3). In the group of 30 patients with more longlasting, relatively quiet, or burned-out disease, atrophic areas that were hypofluorescent on early FA frames and hyperfluorescent on late FA frames tended to be more numerous. There were very often chronic edematous macular changes, but extensive diffuse exudation had mostly resolved.
Indocyanine Green Angiography In patients with active disease (n ⫽ 22), three main ICG angiographic features were noted. The first pattern and most character-
1930
Figure 3. Pseudodelay of retinal arteriovenous circulation. Birdshot chorioretinopathy of 3 years duration. Filling of retinal arteries and capillaries by fluorescein angiography (FA) is complete by 26 seconds (top left) and is followed by diffuse retinal hyperfluorescence (top right). By 40 seconds (bottom left), there still is no FA fluorescence visible in the veins. Indocyanine green angiography (bottom right) shows that arteriovenous circulation is already complete by 23 seconds.
Fardeau et al 䡠 ICG Angiography in Birdshot Chorioretinopathy Table 1. Indocyanine Green Angiography (ICGA) in Birdshot Chorioretinopathy Stage of Disease Active
Intermediate Phase ICGA
Late Phase ICGA
Hypofluorescent dots (100%)
Chronic/treated
Hypofluorescent dots (100%)
佡
FA
Follow-up ICGA Response to Tx
Suspected Lesional Process 1. Full-thickness lesion in choroidal stroma (Fig 1, Fig 5) ⫾ associated with: 2. Choriocapillaris nonperfusion Chorioretinal atrophy (Fig 5) Partial-thickness lesion in choroidal stroma (Fig 4, Fig 6) Leakage from large chroidal vessels (Fig 6) Chronic granulomas or stromal scars ⬇⬎ impaired ICG diffusion intact RPE (Fig 5) Chorioretinal atrophy (Fig 7)
Persistent
Cream-colored areas
Silent or slightly hypofluorescent
Can respond to Tx (Fig 5)
Persistent
Depigmented/atrophic areas Cream-colored areas (often faint)
Hyperfluorescent window defect Silent or slightly hypofluorescent
No response to Tx
Diffuse choroidal hyperfluorescence
NA
NA
Responds to Tx
Persistent
Cream-colored areas round–oval
Silent
No response to Tx
Persistent
Depigmented more irregular forms
Hyperfluorescent window defect
No response to Tx
Erased (isofluorescence) Fuzzy choroidal vessels (100%)
Color Photographs
istic signs shown by ICGA (Table 1), present in all 22 patients with active disease, were hypofluorescent, round– oval dark dots with typically symmetric pattern in both eyes; these varied in number from a few lesions to widespread involvement from the posterior pole up to the equator. Most of these hypofluorescent dark dots were already visible at the arteriovenous (early) phase of ICGA and became more sharply delineated in the intermediate angiographic frames (Figs 1 and 4, middle pictures). The late phase of angiography allowed us to subdivide these hypofluorescent dots into two patterns. Early and intermediate ICG hypofluorescent lesions became either erased (isofluorescent) in the late-phase ICG angiographic frames (Fig 4, bottom pictures) or remained hypofluorescent in the late frames (Fig 1, bottom pictures). Follow-up ICG angiograms after introduction of therapy allowed us to further subdivide the persistent late-phase ICG hypofluorescent areas into lesions responding to therapy (Fig 5) and into lesions remaining hypofluorescent, possibly because of persisting scarred granulomas (not visible on FA) or atrophic lesions (hyperfluorescent on FA due to the window-defect showing total chorioretinal atrophy) (Fig 5, bottom pictures). Most of these round– oval hypofluorescent dots on ICGA were either not seen on FA or, in case of very active disease, were hypofluorescent on FA as well. They were detected in greater number than would correspond to the depigmented areas seen on color and red-free photographs. However, when corresponding areas are rechecked carefully, most of the round ICG hypofluorescent lesions could be distinguished, some of them very faintly, on color and red-free photographs. Their distribution was most dense surrounding the optic nerve and nasal to the disc and occasionally were confluent, especially around the optic nerve (Fig 6). After image analysis by mapping and computerized overlay of the intermediate-phase hypofluorescent dots with red-free photographs, the lesions’ surface area appeared generally smaller than the surface area of the corresponding cream-colored fundus lesions, also visible on the red-free photographs (Fig 6). The second consistent ICG angiographic pattern seen in the group of patients with active disease was diffuse ICG hyperfluorescence predominantly found in the posterior pole in the late phase of angiography, usually in areas where ICG hypofluorescent lesions were present in earlier angiographic phases. This latediffuse choroidal ICG hyperfluorescence was also probably responsible for erasing the ICG hypofluorescent dots present during
Can respond to Tx
Figure 4. Late-diffuse choroidal hyperfluorescence. Nontreated birdshot chorioretinopathy of 3½ years duration. Characteristic patchy “vasculitic” retinal fluorescence on fluorescein angiography (top pictures). Characteristic hypofluorescent areas at the intermediate phase of angiography (middle pictures) replaced in the late phase by late-diffuse choroidal hyperfluorescence having “erased” the hypofluorescent dots visible in the intermediate phase (bottom pictures).
1931
Ophthalmology Volume 106, Number 10, October 1999
Figure 5. Impact of therapy on the hypofluorescent dots. Extensive zones of hypofluorescent dots before treatment (top pictures, top left ⫽ temporal area, top right ⫽ posterior pole) that partially disappear after a 3-day course of intravenous pulse methylprednisolone followed by combined oral prednisone and cyclosporin A treatment (bottom pictures, bottom left ⫽ temporal area, bottom right ⫽ posterior pole).
the intermediate phase of the angiogram (Fig 1, bottom pictures; Fig 4, bottom pictures; Fig 6). The third pattern seen was an alteration of the vascular pattern of the choroid with choroidal vessels appearing fuzzy and indis-
Figure 6. Impact of therapy on fuzzy choroidal vessels and late-diffuse choroidal hyperfluorescence. The left pictures (top and bottom) show diffuse hyperfluorescence and lack of visualization of choroidal vessels. Both signs respond to combined oral prednisone and cyclosporin A therapy (right pictures), with reduction of diffuse hyperfluorescence and more clearly defined choroidal vessels.
1932
Figure 7. Birdshot chorioretinopathy of more than 7 years duration. Clearly visible hypopigmented lesions on blue fundus picture (top). The size of the corresponding hypofluorescent dots usually appears to be smaller than the funduscopic lesions. Note the peripapillary disposition of lesions and the nasal parapapillary atrophic zone, hypofluorescent up to the late phase on indocyanine green angiography (bottom) and hyperfluorescent on fluorescein angiography (not shown).
tinct in the intermediate phase of angiography (Fig 6, left pictures). These vascular alterations usually corresponded to the areas of late-diffuse choroidal hyperfluorescence. In patients with very lightly pigmented fundi, choroidal lesions did not appear on fundus photographs but were very clearly visible on ICGA. In these situations, ICGA was found to have a crucial role in the diagnosis (Fig 1). Furthermore, ICGA was useful in the identification of neovascular membranes in two eyes. In those patients having had pretreatment and post-treatment ICG angiograms (n ⫽ 16), a number of both types of hypofluorescent lesions (present in the intermediate phase and present up to the late phase) resolved under therapy in ten patients but remained unchanged in six patients. However, diffuse choroidal hyperfluorescence and the fuzzy, indistinct appearance of choroidal vessels responded to therapy at least to a certain degree in all 16 cases. These changes were seen as early as 3 days after the initiation of daily intravenous pulses of methylprednisolone in two patients. In the group of 30 patients with more longlasting, controlled, or burned-out disease, round– oval fundus lesions were hypofluorescent at all phases of ICGA and were interpreted to be scarred granulomas (Fig 5, bottom pictures). Atrophic hypofluorescent zones (hyperfluorescent on FA) tended to be more numerous (Fig 7, bottom picture).
Fardeau et al 䡠 ICG Angiography in Birdshot Chorioretinopathy
Discussion Birdshot chorioretinopathy is an immune-mediated inflammatory disease apparently involving only the eyes.1,2 As indicated by its name, the inflammatory process involves both the retina and the choroid.1,2 Retinal involvement has been well-established by FA. The nature and extent of choroidal involvement have been less well-evaluated so far because of limited imaging possibilities of the choroid. In our collective of 52 patients, ICGA showed several signs of choroidal involvement in all 22 patients with active disease, and in 100% of the entire collective, signs of choroidal involvement were seen. By far, the most characteristic ICG angiographic sign was the regularly distributed hypofluorescent dark dots seen in 100% of the patients. Choroidal granulomatous infiltration, which prevented normal choroidal ICG impregnation, most probably was the physiopathogenic explanation for these hypofluorescent dark dots. In the active stage of disease, hypofluorescent dark dots present in the intermediate phase of angiography were seen to either disappear or persist in the late phase of ICGA. Hypofluorescent dots persisting in the late phase were interpreted as full-thickness lesions allowing no ICG diffusion, whereas dots becoming isofluorescent in the late phase were interpreted as partial-thickness lesions progressively surrounded by ICG fluorescence. In the first type of persistent dark dots, additional choriocapillaris nonperfusion may also have been present, which would explain the corresponding hypofluorescence seen on FA in some of the lesions. In chronic disease, most ICG hypofluorescent lesions persisted up to the late phase of angiography and were interpreted as either chronic granulomas or stromal scars when they were not visible by FA or as chorioretinal atrophic areas when they were hyperfluorescent on FA. In the active stage of disease, the two additional ICGA signs, namely fuzzy-appearing vessels and late-diffuse hyperfluorescence, were interpreted as vasculitis with leakage from large choroidal vessels. The latter two signs readily responded to therapy, whereas hypofluorescent dark dots seen in the acute or active stage of disease only partially responded to therapy. This probably indicates that only recent lesions can be made to disappear by therapeutic intervention; more chronic lesions probably are irreversible or leave stromal scars that remain hypofluorescent. The characteristic hypofluorescent dots seen by ICGA usually were not seen on FA. In a case of BC of recent onset showing small, patchy areas of FA hyperfluorescence related to dye leakage from small retinal vessels, comparative analysis of ICGA and FA clearly showed that the patchy (granular) FA hyperfluorescence seen in the retina did not correspond to the ICG hypofluorescent choroidal foci. This clearly indicated that there is a dual, independent involvement of the retina and the choroid and that lesions in one structure (i.e., the retina) usually were not simply secondary to involvement of the adjacent areas in the other structure (i.e., the choroid). Indocyanine green angiography further showed that retinal arteriovenous circulation time was in fact normal in BC. The circulatory pseudodelay originally described by Gass was thus shown not to be a real hemo-
dynamic phenomenon but rather caused only by massive retinal exudation of the small fluorescein molecules very often not reaching a sufficient venous concentration to impregnate the retinal veins.5,6 In two cases, ICGA was useful to identify and determine the extent of inflammatory subretinal neovascular membranes. In cases with hypopigmented fundi, ICGA clearly showed the characteristic multiple, evenly distributed hypofluorescent dark dots that were barely visible on fundus examination. In these situations, ICGA was crucial in establishing the correct diagnosis, whereas in other cases, it was an additional element that helped to establish the right diagnosis. All ICGA features identified in BC were nonspecific for the disease. Similar findings have been identified in other granulomatous diseases involving the choroid, such as Vogt–Koyanagi–Harada disease, sympathetic ophthalmia, posterior sarcoidosis and tuberculosis, and acute posterior multifocal placoid pigment epitheliopathy.7–9,14 –17 For posterior uveitis with predominant choroidal involvement, such as BC, ICGA appears to be an additional useful method to accurately assess ocular inflammatory involvement, thereby determining the extent of choroidal involvement. Furthermore, ICGA may provide information on the pathophysiology of some ocular inflammatory diseases and may also be useful to monitor the effect of therapeutic interventions. Because our patients were investigated by ICGA at different disease and treatment stages, the evolution of ICGA lesions during disease progression was difficult to establish with certainty. The ICG features will now have to be analyzed prospectively from early to late stages of disease to obtain a more accurate semiology in relation to disease stage.
References 1. Ryan SJ, Maumenee AE. Birdshot retinochoroidopathy. Am J Ophthalmol 1980;89:31– 45. 2. LeHoang P, Ryan SJ. Birdshot retinochoroidopathy. In: Pepose JS, Holland GN, Wilhelmus KR, eds. Ocular Infection and Immunity. St. Louis: Mosby, 1996;570 – 8. 3. Priem HA, De Rouck A, De Laey JJ, Bird AC. Electrophysiologic studies in birdshot chorioretinopathy. Am J Ophthalmol 1988;106:430 – 6. 4. de Courten C, Herbort CP. Potential role of computerized visual field testing for the appraisal and follow-up of birdshot chorioretinopathy. Arch Ophthalmol 1998;116:1389 – 91. 5. Gass JDM. Vitiliginous chorioretinitis. Arch Ophthalmol 1981;99:1778 – 87. 6. Guex–Crosier Y, Herbort CP. Prolonged retinal arterio-venous circulation time by fluorescein but not by indocyanine green angiography in birdshot chorioretinopathy. Ocul Immunol Inflamm 1997;5:203– 6. 7. Herbort CP, Borruat FX, de Courten C, Jaccard L. Angiographie au vert d’indocyanine dans les uve´ites poste´rieures. Klin Monatsbl Augenheilkd 1996;208:321– 6. 8. Auer C, Herbort CP. Indocyanine green angiographic features in posterior scleritis. Am J Ophthalmol 1998;126:471– 6.
1933
Ophthalmology Volume 106, Number 10, October 1999 9. Auer C, Bernasconi O, Herbort CP. Toxoplasmic retinochoroiditis: new insights provided by indocyanine green angiography. Am J Ophthalmol 1997;123:131–3. 10. Lim JI, Flower RW. Indocyanine green angiography. Int Ophthalmol Clin 1995;35:59 –70. 11. Herbort CP. Posterior uveitis: new insights provided by indocyanine green angiography [editorial]. Eye 1998;12:757–9. 12. Wolfensberger TJ, Herbort CP. Indocyanine green angiographic features in ocular sarcoidosis. Ophthalmology 1999; 106:285–9. 13. Herbort CP, LeHoang P, Guex–Crosier Y. Schematic interpretation of indocyanine green angiography in posterior uveitis using a standard angiographic protocol. Ophthalmology 1998;105:432– 40.
14. Oshima Y, Harino S, Hara Y, Tano Y. Indocyanine green angiographic findings in Vogt–Koyanagi–Harada disease. Am J Ophthalmol 1996;122:58 – 66. 15. Bernasconi O, Auer C, Zografos L, Herbort CP. Indocyanine green angiographic findings in sympathetic ophthalmia. Graefes Arch Clin Exp Ophthalmol 1998;236:635– 8. 16. Wolfensberger TJ, Piguet B, Herbort CP. Indocyanine green angiographic features in tuberculous chorioretinitis. Am J Ophthalmol 1999;127:350 –3. 17. Howe LJ, Woon H, Graham EM, et al. Choroidal hypoperfusion in acute posterior multifocal placoid pigment epitheliopathy. An indocyanine green angiography study. Ophthalmology 1995;102:790 – 8.
Historical Image Postage stamp commemorating Ludwik Zamenhof.
* Courtesy of the Museum of Ophthalmology, Foundation of the American Academy of Ophthalmology, San Francisco, California.
1934
1928
fuse macular edema, and pseudodelay in arteriovenous fluorescein circulation time.5,6 Inflammatory activity usually is prominent equally in the retina and the choroid.1,2 To assess involvement of the latter, indocyanine green angiography (ICGA) may prove to be very useful as it has in other posterior uveitic conditions involving the choroid.7–9 The absorption peak of indocyanine green (ICG) is situated at ⫾795 nm, and emission of fluorescent light occurs at approximately 830 nm (nearinfrared wavelengths), thus allowing visualization of choroidal fluorescence through the pigmentary epithelium.10 In contrast to fluorescein, ICG is not only highly proteinbound (98%) but binds up to 80% to larger proteins (globulins and alpha-1-lipoproteins). Accordingly, ICG does not leak from normal or only slightly inflamed retinal vessels but leaks unimpaired, albeit slowly, from the fenestrated choriocapillaris, progressively impregnating the choroidal space.11 Because of its particular fluorescence parameters, ICG provides visual access to the choroid and appears to be especially useful to evaluate choroidal involvement of inflammatory disorders.12 The aim of this study was to determine ICGA signs in BC in correlation with clinical signs, disease stage, fundus photographs, and FA findings. We also investigated the possible use of ICGA to monitor both disease progression and effect of immunosuppressive therapy.
Fardeau et al 䡠 ICG Angiography in Birdshot Chorioretinopathy
Patients and Methods Charts of patients with the diagnosis of BC seen in the Department of Ophthalmology of Hoˆpital de la Pitie´-Salpe´trie`re in Paris, the Department of Ophthalmology at Hoˆpital Jules Gonin in Lausanne, and at La Source Eye Center in Lausanne were analyzed.
Inclusion Criteria Patients for whom the diagnosis of BC was verified and who had at least one ICG and fluorescein angiogram performed were included in the study. The diagnosis of BC was based on the clinical and angiographic criteria described in the introduction and by the presence of the HLA-A29 antigen. When necessary, investigations were performed to exclude other possible conditions that could be mistaken for BC, such as sarcoidosis, syphilis, and tuberculosis. Typically, this comprised all or part of the following: complete blood count, erythrocyte sedimentation rate, serum angiotensinconverting enzyme and lysozyme, herpes virus and syphilis serologies, search for autoantibodies and anticardiolipin antibodies, chest x-ray, purified protein derivative skin test, and, occasionally, gallium scan and biopsy of salivary glands. Multifocal choroiditis and multiple evanescent white dot syndrome were excluded on the bases of clinical grounds and angiographic findings.
Treatment The mainstay of therapy consisted of systemic corticosteroid therapy alone in 8 patients, of intravenous immunoglobulins alone in 5 patients, of corticosteroid therapy ⫹ cyclosporin A in 17 patients, and of corticosteroid and/or cyclosporin A and intravenous immunoglobulins in 11 patients; no treatment was given in 11 patients.
Angiographic Procedure A Topcon 1A conventional fundus camera coupled either to an OIS (Opthalmic Imaging System, Sacramento, CA) or an ImageNet (Topcon, Tokyo, Japan) image digitalizing system was used. Conventional fundus camera systems rather than scanning laser ophthalmoscopic systems should be used for uveitis workup because they allow easy panoramic imaging of the whole fundus as well as good reproducibility of angiograms for follow-up examinations and provide good quality late-phase ICG pictures, three very important elements in the angiographic investigation of uveitis. In contrast, the disadvantages of scanning laser ophthalmoscopic videoangiography are partial nonsystematic examination of the fundus, poor visualization of the periphery, and low-quality late ICG pictures. This confocal system allows precise measurement of fluorescence in a given plane of focalization and is therefore very useful when the plane of analysis is well-determined such as in subretinal neovascular membranes. In uveitis, however, lesions may be located from the retinal plane to the scleral plane. Therefore, the global superimposed fluorescence obtained with conventional fundus camera digital angiography is more useful than the focalized images provided by scanning laser ophthalmoscopic systems, which are more difficult to standardize and interpret in routine practice. To exclude autofluorescence, preinjection fluorescence was looked for with the highest flash intensity used for ICGA. At the same time, red-free posterior pole frames were taken. Fluorescein angiography and ICGA usually were performed during the same
session. The procedure typically was initiated with a bolus injection of 50 mg of indocyanine green (Cardiogreen; Becton Dickinson, Cockeysville, MD) diluted in 7.5 ml of a 0.9% salt solution. Early ICGA frames of the posterior pole were taken (to approximately 2–3 minutes; early phase of angiography). Indocyanine green background fluorescence was analyzed in detail at 10 ⫾ 3 minutes after ICG injection (intermediate phase of angiography) by taking frames of the posterior pole and the periphery (a minimum of eight frames of the entire periphery for 360°). At the end of the intermediate phase of ICGA, FA was performed. Early FA frames of the posterior pole were taken up to 2 minutes, followed by 360° panorama imaging of the periphery and late fluorescein posterior pole frames at 10 minutes. Late ICG background fluorescence was analyzed at 35 ⫾ 10 minutes after ICG injection (late phase of angiography) in the same fashion as during the intermediate phase.13
Image Analysis The surface area of hypofluorescent dots in the intermediate phase of ICG angiograms was compared to that of red-free photographs by mapping after computerized overlay. Indocyanine green angiographic findings were related to clinical data, fundus color, green and blue light photographs, and FA findings. The patient history, disease stage, and treatment regimen were other factors considered. In cases with more than one ICG angiogram, evolution of angiographic signs were correlated with clinical evolution and extent of treatment.
Results A total of 52 patients (27 women and 25 men; mean age, 49 ⫾ 7 years at presentation) were included in the study. Twenty-two patients were seen in an active phase of disease, which included both recurrences in insufficiently treated patients and in newly diagnosed patients before treatment. Within this group, both pretreatment and post-treatment angiograms were available on 16 of 22 patients. Thirty patients had their ICG angiograms taken “offtreatment” in a burn-out phase of disease or while they were receiving therapy that was considered to control disease activity. Most of the 52 patients had roughly symmetric clinical involvement, the number of cream-colored areas varying from a few lesions to widespread involvement from the posterior pole to the equator.
Fluorescein Angiography In the active phase of disease, FA hyperfluorescence of the optic disc was present bilaterally in all patients (n ⫽ 22). Diffuse bilateral edema of the posterior pole caused by leakage, particularly along the temporal vascular arcades, was also present in all patients and had a cystoid component in 16 patients (Fig 1). Vasculitis of small retinal vessels was documented in 20 patients, appearing as small FA hyperfluorescent foci scattered over the entire fundus giving a mottled pattern (Figs 1 and 2). These FA hyperfluorescent foci, becoming progressively more fluorescent on late frames of the FA angiogram, indicated retinal inflammatory involvement (Fig 2) because they did not correspond to the creamcolored lesions seen clinically (not shown). The cream-colored fundus lesions were silent (or relatively hypofluorescent) on FA. In seven patients, early FA frames were sufficient to document
1929
Ophthalmology Volume 106, Number 10, October 1999
Figure 2. Dual choroidal and retinal involvement. Newly diagnosed birdshot chorioretinopathy. On fluorescein angiography (FA), the characteristic patchy, mottled hyperfluorescence secondary to vasculitis of small vessels is scattered throughout the fundus. Comparative analysis of indocyanine green angiography (ICGA) and FA shows that the patchy fluorescein hyperfluorescence does not correspond to hypofluorescent areas on ICGA. The latter areas are either silent or hypofluorescent on FA.
Figure 1. Hypofluorescent dots. Birdshot chorioretinopathy with a subacute course of 5 years duration. Fluorescein angiography (top pictures) shows disc and diffuse posterior pole hyperfluorescence (edema). In the periphery (top right), there is typical mottled, patchy hyperfluorescence secondary to vasculitis of small vessels. Indocyanine green angiography (ICGA) shows typical hypofluorescent areas well-apparent in the intermediate phase (middle pictures). Most of the hypofluorescent areas still are present in the late phase (bottom pictures), although some have disappeared in the late phase. Note also diffuse late-phase choroidal hyperfluorescence and fuzziness of choroidal vessels in the intermediate phase (middle pictures). This patient had been followed for 5 years for a uveitis of undetermined origin. Typical pale areas were not present on fundus examination because of an albinotic fundus. Indocyanine green angiography was helpful in making the correct diagnosis.
pseudodelay in retinal arteriovenous fluorescein circulation time, as described by Gass,5 which was not found by ICGA as described earlier6 (Fig 3). In the group of 30 patients with more longlasting, relatively quiet, or burned-out disease, atrophic areas that were hypofluorescent on early FA frames and hyperfluorescent on late FA frames tended to be more numerous. There were very often chronic edematous macular changes, but extensive diffuse exudation had mostly resolved.
Indocyanine Green Angiography In patients with active disease (n ⫽ 22), three main ICG angiographic features were noted. The first pattern and most character-
1930
Figure 3. Pseudodelay of retinal arteriovenous circulation. Birdshot chorioretinopathy of 3 years duration. Filling of retinal arteries and capillaries by fluorescein angiography (FA) is complete by 26 seconds (top left) and is followed by diffuse retinal hyperfluorescence (top right). By 40 seconds (bottom left), there still is no FA fluorescence visible in the veins. Indocyanine green angiography (bottom right) shows that arteriovenous circulation is already complete by 23 seconds.
Fardeau et al 䡠 ICG Angiography in Birdshot Chorioretinopathy Table 1. Indocyanine Green Angiography (ICGA) in Birdshot Chorioretinopathy Stage of Disease Active
Intermediate Phase ICGA
Late Phase ICGA
Hypofluorescent dots (100%)
Chronic/treated
Hypofluorescent dots (100%)
佡
FA
Follow-up ICGA Response to Tx
Suspected Lesional Process 1. Full-thickness lesion in choroidal stroma (Fig 1, Fig 5) ⫾ associated with: 2. Choriocapillaris nonperfusion Chorioretinal atrophy (Fig 5) Partial-thickness lesion in choroidal stroma (Fig 4, Fig 6) Leakage from large chroidal vessels (Fig 6) Chronic granulomas or stromal scars ⬇⬎ impaired ICG diffusion intact RPE (Fig 5) Chorioretinal atrophy (Fig 7)
Persistent
Cream-colored areas
Silent or slightly hypofluorescent
Can respond to Tx (Fig 5)
Persistent
Depigmented/atrophic areas Cream-colored areas (often faint)
Hyperfluorescent window defect Silent or slightly hypofluorescent
No response to Tx
Diffuse choroidal hyperfluorescence
NA
NA
Responds to Tx
Persistent
Cream-colored areas round–oval
Silent
No response to Tx
Persistent
Depigmented more irregular forms
Hyperfluorescent window defect
No response to Tx
Erased (isofluorescence) Fuzzy choroidal vessels (100%)
Color Photographs
istic signs shown by ICGA (Table 1), present in all 22 patients with active disease, were hypofluorescent, round– oval dark dots with typically symmetric pattern in both eyes; these varied in number from a few lesions to widespread involvement from the posterior pole up to the equator. Most of these hypofluorescent dark dots were already visible at the arteriovenous (early) phase of ICGA and became more sharply delineated in the intermediate angiographic frames (Figs 1 and 4, middle pictures). The late phase of angiography allowed us to subdivide these hypofluorescent dots into two patterns. Early and intermediate ICG hypofluorescent lesions became either erased (isofluorescent) in the late-phase ICG angiographic frames (Fig 4, bottom pictures) or remained hypofluorescent in the late frames (Fig 1, bottom pictures). Follow-up ICG angiograms after introduction of therapy allowed us to further subdivide the persistent late-phase ICG hypofluorescent areas into lesions responding to therapy (Fig 5) and into lesions remaining hypofluorescent, possibly because of persisting scarred granulomas (not visible on FA) or atrophic lesions (hyperfluorescent on FA due to the window-defect showing total chorioretinal atrophy) (Fig 5, bottom pictures). Most of these round– oval hypofluorescent dots on ICGA were either not seen on FA or, in case of very active disease, were hypofluorescent on FA as well. They were detected in greater number than would correspond to the depigmented areas seen on color and red-free photographs. However, when corresponding areas are rechecked carefully, most of the round ICG hypofluorescent lesions could be distinguished, some of them very faintly, on color and red-free photographs. Their distribution was most dense surrounding the optic nerve and nasal to the disc and occasionally were confluent, especially around the optic nerve (Fig 6). After image analysis by mapping and computerized overlay of the intermediate-phase hypofluorescent dots with red-free photographs, the lesions’ surface area appeared generally smaller than the surface area of the corresponding cream-colored fundus lesions, also visible on the red-free photographs (Fig 6). The second consistent ICG angiographic pattern seen in the group of patients with active disease was diffuse ICG hyperfluorescence predominantly found in the posterior pole in the late phase of angiography, usually in areas where ICG hypofluorescent lesions were present in earlier angiographic phases. This latediffuse choroidal ICG hyperfluorescence was also probably responsible for erasing the ICG hypofluorescent dots present during
Can respond to Tx
Figure 4. Late-diffuse choroidal hyperfluorescence. Nontreated birdshot chorioretinopathy of 3½ years duration. Characteristic patchy “vasculitic” retinal fluorescence on fluorescein angiography (top pictures). Characteristic hypofluorescent areas at the intermediate phase of angiography (middle pictures) replaced in the late phase by late-diffuse choroidal hyperfluorescence having “erased” the hypofluorescent dots visible in the intermediate phase (bottom pictures).
1931
Ophthalmology Volume 106, Number 10, October 1999
Figure 5. Impact of therapy on the hypofluorescent dots. Extensive zones of hypofluorescent dots before treatment (top pictures, top left ⫽ temporal area, top right ⫽ posterior pole) that partially disappear after a 3-day course of intravenous pulse methylprednisolone followed by combined oral prednisone and cyclosporin A treatment (bottom pictures, bottom left ⫽ temporal area, bottom right ⫽ posterior pole).
the intermediate phase of the angiogram (Fig 1, bottom pictures; Fig 4, bottom pictures; Fig 6). The third pattern seen was an alteration of the vascular pattern of the choroid with choroidal vessels appearing fuzzy and indis-
Figure 6. Impact of therapy on fuzzy choroidal vessels and late-diffuse choroidal hyperfluorescence. The left pictures (top and bottom) show diffuse hyperfluorescence and lack of visualization of choroidal vessels. Both signs respond to combined oral prednisone and cyclosporin A therapy (right pictures), with reduction of diffuse hyperfluorescence and more clearly defined choroidal vessels.
1932
Figure 7. Birdshot chorioretinopathy of more than 7 years duration. Clearly visible hypopigmented lesions on blue fundus picture (top). The size of the corresponding hypofluorescent dots usually appears to be smaller than the funduscopic lesions. Note the peripapillary disposition of lesions and the nasal parapapillary atrophic zone, hypofluorescent up to the late phase on indocyanine green angiography (bottom) and hyperfluorescent on fluorescein angiography (not shown).
tinct in the intermediate phase of angiography (Fig 6, left pictures). These vascular alterations usually corresponded to the areas of late-diffuse choroidal hyperfluorescence. In patients with very lightly pigmented fundi, choroidal lesions did not appear on fundus photographs but were very clearly visible on ICGA. In these situations, ICGA was found to have a crucial role in the diagnosis (Fig 1). Furthermore, ICGA was useful in the identification of neovascular membranes in two eyes. In those patients having had pretreatment and post-treatment ICG angiograms (n ⫽ 16), a number of both types of hypofluorescent lesions (present in the intermediate phase and present up to the late phase) resolved under therapy in ten patients but remained unchanged in six patients. However, diffuse choroidal hyperfluorescence and the fuzzy, indistinct appearance of choroidal vessels responded to therapy at least to a certain degree in all 16 cases. These changes were seen as early as 3 days after the initiation of daily intravenous pulses of methylprednisolone in two patients. In the group of 30 patients with more longlasting, controlled, or burned-out disease, round– oval fundus lesions were hypofluorescent at all phases of ICGA and were interpreted to be scarred granulomas (Fig 5, bottom pictures). Atrophic hypofluorescent zones (hyperfluorescent on FA) tended to be more numerous (Fig 7, bottom picture).
Fardeau et al 䡠 ICG Angiography in Birdshot Chorioretinopathy
Discussion Birdshot chorioretinopathy is an immune-mediated inflammatory disease apparently involving only the eyes.1,2 As indicated by its name, the inflammatory process involves both the retina and the choroid.1,2 Retinal involvement has been well-established by FA. The nature and extent of choroidal involvement have been less well-evaluated so far because of limited imaging possibilities of the choroid. In our collective of 52 patients, ICGA showed several signs of choroidal involvement in all 22 patients with active disease, and in 100% of the entire collective, signs of choroidal involvement were seen. By far, the most characteristic ICG angiographic sign was the regularly distributed hypofluorescent dark dots seen in 100% of the patients. Choroidal granulomatous infiltration, which prevented normal choroidal ICG impregnation, most probably was the physiopathogenic explanation for these hypofluorescent dark dots. In the active stage of disease, hypofluorescent dark dots present in the intermediate phase of angiography were seen to either disappear or persist in the late phase of ICGA. Hypofluorescent dots persisting in the late phase were interpreted as full-thickness lesions allowing no ICG diffusion, whereas dots becoming isofluorescent in the late phase were interpreted as partial-thickness lesions progressively surrounded by ICG fluorescence. In the first type of persistent dark dots, additional choriocapillaris nonperfusion may also have been present, which would explain the corresponding hypofluorescence seen on FA in some of the lesions. In chronic disease, most ICG hypofluorescent lesions persisted up to the late phase of angiography and were interpreted as either chronic granulomas or stromal scars when they were not visible by FA or as chorioretinal atrophic areas when they were hyperfluorescent on FA. In the active stage of disease, the two additional ICGA signs, namely fuzzy-appearing vessels and late-diffuse hyperfluorescence, were interpreted as vasculitis with leakage from large choroidal vessels. The latter two signs readily responded to therapy, whereas hypofluorescent dark dots seen in the acute or active stage of disease only partially responded to therapy. This probably indicates that only recent lesions can be made to disappear by therapeutic intervention; more chronic lesions probably are irreversible or leave stromal scars that remain hypofluorescent. The characteristic hypofluorescent dots seen by ICGA usually were not seen on FA. In a case of BC of recent onset showing small, patchy areas of FA hyperfluorescence related to dye leakage from small retinal vessels, comparative analysis of ICGA and FA clearly showed that the patchy (granular) FA hyperfluorescence seen in the retina did not correspond to the ICG hypofluorescent choroidal foci. This clearly indicated that there is a dual, independent involvement of the retina and the choroid and that lesions in one structure (i.e., the retina) usually were not simply secondary to involvement of the adjacent areas in the other structure (i.e., the choroid). Indocyanine green angiography further showed that retinal arteriovenous circulation time was in fact normal in BC. The circulatory pseudodelay originally described by Gass was thus shown not to be a real hemo-
dynamic phenomenon but rather caused only by massive retinal exudation of the small fluorescein molecules very often not reaching a sufficient venous concentration to impregnate the retinal veins.5,6 In two cases, ICGA was useful to identify and determine the extent of inflammatory subretinal neovascular membranes. In cases with hypopigmented fundi, ICGA clearly showed the characteristic multiple, evenly distributed hypofluorescent dark dots that were barely visible on fundus examination. In these situations, ICGA was crucial in establishing the correct diagnosis, whereas in other cases, it was an additional element that helped to establish the right diagnosis. All ICGA features identified in BC were nonspecific for the disease. Similar findings have been identified in other granulomatous diseases involving the choroid, such as Vogt–Koyanagi–Harada disease, sympathetic ophthalmia, posterior sarcoidosis and tuberculosis, and acute posterior multifocal placoid pigment epitheliopathy.7–9,14 –17 For posterior uveitis with predominant choroidal involvement, such as BC, ICGA appears to be an additional useful method to accurately assess ocular inflammatory involvement, thereby determining the extent of choroidal involvement. Furthermore, ICGA may provide information on the pathophysiology of some ocular inflammatory diseases and may also be useful to monitor the effect of therapeutic interventions. Because our patients were investigated by ICGA at different disease and treatment stages, the evolution of ICGA lesions during disease progression was difficult to establish with certainty. The ICG features will now have to be analyzed prospectively from early to late stages of disease to obtain a more accurate semiology in relation to disease stage.
References 1. Ryan SJ, Maumenee AE. Birdshot retinochoroidopathy. Am J Ophthalmol 1980;89:31– 45. 2. LeHoang P, Ryan SJ. Birdshot retinochoroidopathy. In: Pepose JS, Holland GN, Wilhelmus KR, eds. Ocular Infection and Immunity. St. Louis: Mosby, 1996;570 – 8. 3. Priem HA, De Rouck A, De Laey JJ, Bird AC. Electrophysiologic studies in birdshot chorioretinopathy. Am J Ophthalmol 1988;106:430 – 6. 4. de Courten C, Herbort CP. Potential role of computerized visual field testing for the appraisal and follow-up of birdshot chorioretinopathy. Arch Ophthalmol 1998;116:1389 – 91. 5. Gass JDM. Vitiliginous chorioretinitis. Arch Ophthalmol 1981;99:1778 – 87. 6. Guex–Crosier Y, Herbort CP. Prolonged retinal arterio-venous circulation time by fluorescein but not by indocyanine green angiography in birdshot chorioretinopathy. Ocul Immunol Inflamm 1997;5:203– 6. 7. Herbort CP, Borruat FX, de Courten C, Jaccard L. Angiographie au vert d’indocyanine dans les uve´ites poste´rieures. Klin Monatsbl Augenheilkd 1996;208:321– 6. 8. Auer C, Herbort CP. Indocyanine green angiographic features in posterior scleritis. Am J Ophthalmol 1998;126:471– 6.
1933
Ophthalmology Volume 106, Number 10, October 1999 9. Auer C, Bernasconi O, Herbort CP. Toxoplasmic retinochoroiditis: new insights provided by indocyanine green angiography. Am J Ophthalmol 1997;123:131–3. 10. Lim JI, Flower RW. Indocyanine green angiography. Int Ophthalmol Clin 1995;35:59 –70. 11. Herbort CP. Posterior uveitis: new insights provided by indocyanine green angiography [editorial]. Eye 1998;12:757–9. 12. Wolfensberger TJ, Herbort CP. Indocyanine green angiographic features in ocular sarcoidosis. Ophthalmology 1999; 106:285–9. 13. Herbort CP, LeHoang P, Guex–Crosier Y. Schematic interpretation of indocyanine green angiography in posterior uveitis using a standard angiographic protocol. Ophthalmology 1998;105:432– 40.
14. Oshima Y, Harino S, Hara Y, Tano Y. Indocyanine green angiographic findings in Vogt–Koyanagi–Harada disease. Am J Ophthalmol 1996;122:58 – 66. 15. Bernasconi O, Auer C, Zografos L, Herbort CP. Indocyanine green angiographic findings in sympathetic ophthalmia. Graefes Arch Clin Exp Ophthalmol 1998;236:635– 8. 16. Wolfensberger TJ, Piguet B, Herbort CP. Indocyanine green angiographic features in tuberculous chorioretinitis. Am J Ophthalmol 1999;127:350 –3. 17. Howe LJ, Woon H, Graham EM, et al. Choroidal hypoperfusion in acute posterior multifocal placoid pigment epitheliopathy. An indocyanine green angiography study. Ophthalmology 1995;102:790 – 8.
Historical Image Postage stamp commemorating Ludwik Zamenhof.
* Courtesy of the Museum of Ophthalmology, Foundation of the American Academy of Ophthalmology, San Francisco, California.
1934